Sentence Similarity
sentence-transformers
Safetensors
Transformers
qwen3_vl
image-text-to-text
multimodal embedding
qwen
embedding
Instructions to use Qwen/Qwen3-VL-Embedding-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use Qwen/Qwen3-VL-Embedding-2B with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("Qwen/Qwen3-VL-Embedding-2B") sentences = [ "That is a happy person", "That is a happy dog", "That is a very happy person", "Today is a sunny day" ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [4, 4] - Transformers
How to use Qwen/Qwen3-VL-Embedding-2B with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForImageTextToText processor = AutoProcessor.from_pretrained("Qwen/Qwen3-VL-Embedding-2B") model = AutoModelForImageTextToText.from_pretrained("Qwen/Qwen3-VL-Embedding-2B") - Notebooks
- Google Colab
- Kaggle
Instruction issue
#18
by hh1212 - opened
Thanks for your great work!
I would like to know how to specifically configure the instructions in your workflow. For instance, in a retrieval task: Should I set the instruction for extracting gallery embeddings to: 'Represent this document for retrieval: '? And for the downstream query (extracting query embeddings), should I set it to: 'Represent this query for searching relevant documents: '?
I look forward to hearing from you.